Chapter 5

CAPACITY PLANNING AND AGGREGATE PRODUCTION PLANNING

5.1 Capacity Planning
5.2 Rough Cut Capacity Planning
5.3 Capacity Requirement Planning
5.4 Input / Output Control
5.5 Aggregate Production Planning
5.6 Aggregate Planning for Services

Chapter Objectives

By the end of the chapter students should be in position to:

1. Define capacity and other related terms

2. List and explain capacity expansion strategies
3. List and explain capacity levels
4. Evaluate capacity alternatives

Hierarchical Planning Process

By determining a strategy for meeting and managing demand, aggregate planning provides a
framework within which shorter-term production and capacity decisions can be made. In
production planning, the next level of detail is a master production schedule, in which
weekly (not monthly or quarterly) production plans are specified by individual final product
(not product line). At another level of detail, material requirements planning deals with the
production of the components that go into the final products. Shop floor scheduling schedules
the manufacturing operations required to make each component.

In capacity planning, we might develop a resource requirements plan, to verify that an

aggregate production plan is doable, and a rough-cut capacity plan as a quick check to see if
the master production schedule is feasible. One level down, a much more detailed capacity
requirements plan that matches the factory's machine and labor resources to the material
requirements plan is developed. Finally, the input/output control is used to monitor the
production that takes place at individual machines or work centers. At each level, decisions
are made within the parameters set by the higher-level decisions. The process of moving
from the aggregate plan to the next level down is called disaggregation. Each level of
production and capacity planning is shown in figure 5.1.

The five levels of capacity planning activities range from large aggregations of capacity for
long time periods to very detailed machine scheduling, for an hour or shorter time interval.

Most of the chapter discussion will focus on rough cut capacity planning procedures (for
medium range capacity planning) and on the technique called capacity requirements planning
(short range capacity requirements determined for individual work centers).
Figure 5.1

5.1 Capacity Planning

Capacity planning is a long-term strategic decision that establishes a firm's overall level of
resources. It extends over a time horizon long enough to obtain those resources--usually a
year or more for building new facilities or acquiring new businesses. Capacity decisions
affect product lead times, customer responsiveness, operating costs, and a firm's ability to
compete. Inadequate capacity can lead to loss of customers, escalating work in process, limit
growth and frustrated manufacturing personnel who will quickly turn back to the informal
system to solve problems. Excess capacity can drain a company's resources and prevent
investments in more lucrative ventures. When to increase capacity and how much to increase
capacity are critical decisions.

The common capacity expansion strategies are namely: capacity lead strategy, capacity lag
strategy, average capacity strategy and incremental one step expansion strategy.
Figure 5.2

Figures 5.2 (a), (b), and (c) show three basic strategies for the timing of capacity expansion
in relation to a steady growth in demand.

 Capacity lead strategy. Capacity is expanded in anticipation of demand growth. This

aggressive strategy is used to lure customers from competitors who are capacity
constrained or to gain a foothold in a rapidly expanding market.
 Capacity lag strategy. Capacity is increased after an increase in demand has been
documented. This conservative strategy produces a higher return on investment but
may lose customers in the process. It is used in industries with standard products and
cost-based or weak competition. The strategy assumes that lost customers will return
from competitors after capacity has been expanded.
 Average capacity strategy. Capacity is expanded to coincide with average expected
demand. This is a moderate strategy in which managers are certain they will be able
to sell at least some portion of the additional output.

Consider higher education's strategy in preparing for a tripling of the state's college-bound
population in the next decade. An established university, guaranteed applicants even in lean
years, may follow a capacity lag strategy. A young university might lead capacity expansion
in hopes of capturing those students not admitted to the more established universities. A
community college may choose the average capacity strategy to fulfill its mission of
educating the state's youth but with little risk.

How much to increase capacity depends on (1) the volume and certainty of anticipated
demand; (2) strategic objectives in terms of growth, customer service, and competition; and
(3) the costs of expansion and operation.

Capacity can be increased incrementally or in one large step as shown in Figure 5.2(d).
Incremental expansion is less risky but more costly. An attractive alternative to expanding
capacity is outsourcing, in which suppliers absorb the risk of demand uncertainty.
The best operating level for a facility is the percent of capacity utilization that minimizes
average unit cost. Rarely is the best operating level at 100 percent of capacity--at higher
levels of utilization, productivity slows and things start to go wrong. Average capacity
utilization differs by industry. An industry with an 80 percent average utilization would have
a 20 percent capacity cushion for unexpected surges in demand or temporary work
stoppages. Large capacity cushions are common in industries where demand is highly
variable, resource flexibility is low, and customer service is important. Utilities, for example,
maintain a 20 percent capacity cushion. Capital-intensive industries with less flexibility and
higher costs maintain cushions under 10 percent. Airlines maintain a negative cushion--
overbooking is a common practice!

Figure 5.3

Figure 5.3 shows the best operating level--in this case, the optimal occupancy rate--for three
different size hotels. Of the three alternatives, the 500-room hotel has the lowest average unit
cost. This is the point where the economies of scale have reached their peak and the
diseconomies of scale have not yet begun.

High levels of output tend to cost less per unit. Called economies of scale, this holds true
when:

 Fixed costs can be spread over a larger number of units,

 Production costs do not increase linearly with output levels,
 Quantity discounts are available for material purchases, and
 Production efficiency increases as workers gain experience.

The electronics industry provides a good case example of economies of scale. The average
cost per chip placement for printed circuit-board assembly is 32 cents in factories with a
volume of 25 million placements, 15 cents in factories with 200 million placements, and only
10 cents in factories with 800 million placements.
Economies of scale do not continue indefinitely. Above a certain level of output,
diseconomies of scale can occur. Overtaxed machines and material handling equipment
break down, service time slows, quality suffers requiring more rework, labor costs increase
with overtime, and coordination and management activities become difficult. In addition, if
customer preferences suddenly change, high-volume production can leave a firm with
unusable inventory and excess capacity.

Long-term capacity decisions concerning the number of facilities and facility size provide the
framework for making more intermediate-term capacity decisions--such as inventory
policies, production rates, and staffing levels. These decisions are collectively known as
aggregate production planning or just plain aggregate planning.

5.2 Rough-Cut Capacity Planning

Rough-cut capacity planning checks whether critical resources are available to support the
preliminary master schedule. A resource bill shows the time required for individual items on
a critical resource. Once the preliminary MPSs are made, they must be checked against the
available capacity. This process is called rough-cut capacity planning; its purpose is to check
whether critical resources are available to support the master production schedule.

A key concern in our ability to meet the master schedule is the existence of bottlenecks. You
may not know where your bottleneck work centers are, but you can be sure you have them.
Suppose you have inventory piled up at each work center. Something is definitely wrong. At
your bottleneck work center, you want to have a backlog of work so that it’s never out of
material, but everything upstream of the bottleneck center shouldn’t have backlog; they
should be running with excess capacity. It may be necessary to take work orders off the shop
floor and remove WIP from the work centers.

If that’s the case, you need to do a rough-cut capacity check. Check those key work centers
and check your resource bill. Make sure you can build what you need to. You’ll probably
find that your demand and your resources are out of balance. Once you find out that they are
out of balance, the next thing you’re going to have to do is change something.

You’re going to have to increase capacity if you need more capacity; you’re going to have to
get more material if you need more material; you’re going to have to shift skills if that’s what
you need. You may also have to change the master schedule. What? Change the master
schedule? Of course! The master schedule is simply a tool.

5.3 Capacity Requirements Planning (CRP)

Capacity requirements planning (CRP) is a computerized system that projects the load
from a given material plan onto the capacity of a system and identifies under loads and
overloads. It is then up to the MRP planner to level the load--smooth out the resource
requirements so that capacity constraints are not violated. This can be accomplished by
shifting requirements, reducing requirements, or temporarily expanding capacity.

There are three major inputs to CRP, as shown in figure 5.4

Figure 5.4

 The planned order releases from the MRP process;

 A routing file, which specifies which machines are required to complete an order
from the MRP plan, in what order the operations are to be conducted, and the length
of time each operation should take; and
 An open orders file, which contains information on the status of jobs that have
already been released to the shop but have not yet been completed.

With this information, CRP can produce a load profile for each machine or work center in
the shop. The load profile compares released orders and planned orders with work center
capacity.

Capacity, usually expressed as standard machine hours or labor hours, is calculated as

follows:

Capacity = (no. machines or workers) × (no. shifts) × (utilization) × (efficiency)

Utilization refers to the percentage of available working time that a worker actually works or
a machine actually runs. Scheduled maintenance, lunch breaks, and setup time are examples
of activities that reduce actual working time. Efficiency refers to how well a machine or
worker performs compared to a standard output level. Standards can be based on past records
of performance or can be developed from the work-measurement techniques. An efficiency
of 100 percent is considered normal or standard performance, 125 percent is above normal,
and 90 percent is below normal. Efficiency is also dependent on product mix. Some orders
obviously will take longer than others to process, and some machines or workers may be
better at processing certain types of orders.

Load is the standard hours of work (or equivalent units of production) assigned to a
production facility. After load and capacity have been determined, a load percent can be
calculated as
Centers loaded above 100 percent will not be able to complete the scheduled work without
some adjustment in capacity or reduction in load.

EXAMPLE
Determining Loads and Capacities
5.1

Copy Courier is a fledging copy center in downtown Richmond run by two college
students. Currently, the equipment consists of two high-speed copiers that can be
operated by one operator. If the students work alone, it is conceivable that two shifts per
day can be staffed. The students each work 8 hours a day, 5 days a week. They do not
take breaks during the day, but they do allow themselves 30 minutes for lunch or dinner.
In addition, they service the machines for about 30 minutes at the beginning of each shift.
The time required to set up for each order varies by the type of paper, use of color,
number of copies, and so on. Estimates of setup time are kept with each order. Since the
machines are new, their efficiency is estimated at 100 percent.

Due to extensive advertising and new customer incentives, orders have been pouring in.
The students need help determining the capacity of their operation and the current load on
their facility. Use the following information to calculate the normal daily capacity of
Copy Courier and to project next Monday's load profile and load percent:

SOLUTION:

The machines and/or operators at Copy Courier are out of service for 1 hour each shift for
maintenance and lunch. Utilization is, thus, 7/8, or 87.5 percent. Daily copy shop
capacity is

2 machines × 2 shifts ×8 hours/shift ×100% efficiency × 87.5% utilization

= 28 hours or 1,680 minutes

The projected load for Monday of next week is as follows:

 Copy Courier is loaded 42 percent over capacity next Monday.

Increasing utilization (even to 100 percent) would not be sufficient to get the work done.
To complete the customer orders on time, another shift could be added (i.e., another
person hired). With this adjustment, the copy shop's daily capacity would increase to

2 machines × 3 shifts × 8 hours/shift × 100% efficiency × 87.5% utilization

= 42 hours or 2,520 minutes

The revised load percent is:

In the future, Copy Courier should determine if it has enough capacity to complete a job
by the customer's requested due date before the job is accepted.

Load profiles can be displayed graphically, as shown in figure 5.5. The normal capacity of
machine 32A is 40 hours per week. We can see that the machine is underloaded in periods 1,
5, and 6, and overloaded in periods 2, 3, and 4.

Figure 5.5: Load profile for machine 32A

Under-loaded conditions can be leveled by:

1. Acquiring more work;

2. Pulling work ahead that is scheduled for later periods; or
3. Reducing normal capacity.
Additional work can be acquired by transferring similar work from other machines in the
same shop that are near or over capacity, by making components in-house that are normally
purchased from outside suppliers, or by seeking work from outside sources. Pulling work
ahead seems like a quick and easy alternative to alleviate both under-loads and overloads.
However, we must remember that the MRP plan was devised based on an interrelated
product structure, so the feasibility of scheduling work in an earlier time period is contingent
on the availability of required materials or components. In addition, work completed prior to
its due date must be stored in inventory and thus incurs a holding cost. When work is shifted
to other time periods, the MRP plan should be rerun to check the feasibility of the proposed
schedule.

If an under loaded condition continues for some time, reducing the size of the workforce may
be necessary. Smaller under loads can be handled by reducing the length of the working day
or workweek, by scheduling idled workers for training sessions or vacations, or by
transferring workers to other positions at machine centers or departments where overloads
are occurring.

Overloaded conditions are the primary concern of the MRP planner because an overloaded
schedule left unchecked cannot possibly be completed as planned. Overloads can be reduced
by:

1. Eliminating unnecessary requirements;

2. Rerouting jobs to alternative machines or work centers;
3. Splitting lots between two or more machines;
4. Increasing normal capacity;
5. Subcontracting;
6. Increasing the efficiency of the operation;
7. Pushing work back to later time periods; or
8. Revising the master schedule.

Some capacity problems are generated from an MRP plan that includes lot sizes, safety stock,
or unsubstantiated requirements for service parts or forecasted demand. To verify that a
capacity overload is caused by "real" need, the planner might examine the MRP matrices of
the items processed through a machine center during an overloaded period as well as the
matrices of the parents of those items processed, all the way up the product structure to the
master schedule. Or, the MRP system could be rerun with lot sizes temporarily set to one and
safety stock to zero to see if the capacity problem is eliminated.

MRP systems assume that an entire lot of goods is processed at one machine. Given the job
shop environment in which most MRP systems are installed, there are usually several
machines that can perform the same job (although perhaps not as efficiently). With CRP,
load profiles are determined with jobs assigned to the preferred machine first, but when
capacity problems occur, jobs can certainly be reassigned to alternate machines. In addition,
if two or more similar machines are available at the same time, it may be possible to split a
batch--that is, assign part of an order to one machine and the remainder to another machine.
 EXAMPLE
Splitting Orders
5.2

Duffy's Machine Shop has a shortage of lathes. Next week's schedule loaded the lathe
department 125 percent. Management's usual response is to schedule overtime, but the
company is in a tight financial bind and wants to evaluate other options. The shop
supervisor, who has been reading about methods for reducing processing time, suggests
something called order splitting.

It turns out that some of the lathe work can actually be performed on a milling machine, but
it is rarely done that way because the process takes longer and the setup is more involved.
Setup time for lathes averages 30 minutes, whereas setup for milling machines averages 45
minutes. Processing time per piece is 5 minutes on the lathe, compared to 10 minutes on the
milling machine. Management is wondering what the effect would be of producing an entire
order of 100 pieces on the lathe or splitting the order in half between the lathe and milling
machine. Further, if the objective is to complete the order as soon as possible, is there an
optimum split between the two types of machines?

SOLUTION:

If the order were processed on lathe alone, it would take

If the order were equally split between lathe and milling machine, the processing time at
each machine would be:

Assuming that the lathe and milling machine are run simultaneously, the completion time
for the entire order of 100 units is calculated by determining the completion time at each
machine and taking the largest number, in this case, 545 minutes. Thus, if the order were
equally split between the two machines, it would actually take longer to complete.

Determining the optimal split between machines requires algebra. We want the machines to
finish processing at the same time, so we need to equate the processing-time equations for
each machine and solve for the optimum number of units processed. If we let x represent
the number of units processed on the lathe, then (100  x) is the number of units processed
on the milling machine.
Thus, the optimal split would process 68 units on the lathe and 32 units on the milling
machine. Completion time for the optimal split is calculated as follows:

By splitting the order, it can be completed in 370 minutes, versus the 530 minutes on the
lathe alone. That is a 39 percent reduction in processing time. Applied to the weekly
demand, the 25 percent overload could be alleviated by splitting orders.

Normal capacity can be increased by adding extra hours to the workday, extra days to the
workweek, or extra shifts. Temporary overloads are usually handled with overtime. More
extensive overloads may require hiring additional workers. Work can also be outsourced.

Improving the efficiency of an operation increases its capacity. Assigning the most efficient
workers to an overloaded machine, improving the operating procedures or tools, or
decreasing the percentage of items that need to be reworked or scrapped increases efficiency
and allows more items to be processed in the same amount of time. Because output increases
with the same amount of input, productivity increases. This is especially useful for alleviating
chronic overloads at bottleneck operations, but it does take time to put into effect.

If later time periods are under loaded, it may be possible to push work back to those periods,
so that the work is completed but later than originally scheduled. There are two problems
with this approach. First, postponing some jobs could throw the entire schedule off, meaning
customers will not receive the goods when promised. This could involve a penalty for late
delivery, loss of an order, or loss of a customer. Second, filling up the later time periods may
preclude accepting new orders in those periods. It is normal for time periods further in the
future to be under loaded. As these periods draw nearer, customer orders accelerate and begin
taking up more of the system's capacity.

If all the preceding approaches to remedying overloads have been tried, but an overload still
exists, the only option is to revise the master schedule. That means some customer will not
receive goods as previously promised. The planner, in conjunction with someone from
marketing, should determine which customer has the lowest priority and whether its order
should be postponed or canceled.

There are cost consequences associated with each of these alternatives, but there is usually no
attempt to derive an optimum solution. More than likely, the MRP planner will use the
options that produce a feasible solution quickly. In many manufacturing environments, new
customer orders arrive daily, and feasible MRP plans can become infeasible overnight.

Figure 5.6

Figure 5.6 shows one possible remedy for the overloads shown in Figure 5.5. Ten hours of
work are pulled ahead from period 2 to period 1. Ten hours of overtime are assigned in
period 2. An entire 40-hour shift is added in period 3. Ten hours of work from period 4 are
pushed back to period 5, and 20 hours are pushed back to period 6.

CRP identifies capacity problems, but the planner solves the problems. With experience, the
task of shifting work and leveling loads is not as formidable as it appears. However, it is
helpful if the initial load profile is as accurate as possible and if previous planning stages (i.e.
aggregate production planning and master production scheduling) have considered capacity
constraints. Some companies formalize capacity planning at each stage of production
planning. Resource requirements planning is associated with an aggregate production plan,
and rough-cut capacity planning is performed prior to the approval of a master schedule.
Capacity requirements planning may still be performed on the material requirements plan,
but its role is to fine-tune existing resources, rather than to find or develop new resources.

Once the feasibility of an MRP plan has been verified by CRP, the plan can be executed by
releasing orders in the time periods indicated. Early MRP systems had no mechanism for
monitoring the success of their plans. Today's MRP systems include elaborate capacity and
reporting modules for scheduling and monitoring daily work requirements.

5.4 Input / Output Control

Input/output (I/O) control monitors the input to and output from each work center. Prior to
such analysis, it was common to examine only the output from a work center and to compare
the actual output with the output planned in the shop schedule. Using that approach in a job
shop environment in which the performance of different work centers is interrelated may
result in erroneous conclusions about the source of a problem. Reduced output at one point in
the production process may be caused by problems at the current work center, but it may also
be caused by problems at previous work centers that feed the current work center. Thus, to
identify more clearly the source of a problem, the input to a work center must be compared
with the planned input, and the output must be compared with the planned output. Deviations
between planned and actual values are calculated, and their cumulative effects are observed.
The resulting backlog or queue size is monitored to ensure that it stays within a manageable
range.

The input rate to a work center can be controlled only for the initial operations of a job.
These first work centers are often called gateway work centers, because the majority of jobs
must pass through them before subsequent operations are performed. Input to later
operations, performed at downstream work centers, is difficult to control because it is a
function of how well the rest of the shop is operating--that is, where queues are forming and
how smoothly jobs are progressing through the system. The deviation of planned to actual
input for downstream work centers can be minimized by controlling the output rates of
feeding work centers. The use of input/output reports can best be illustrated with an example.

EXAMPLE
Input/Output Control
5.4

The following information has been compiled in an input/output report for work center 5.
Complete the report and interpret the results.

SOLUTION:

The input/output report has planned a level production of 75 units per period for work
center 5. This is to be accomplished by working off the backlog of work and steadily
increasing the input of work.

The report is completed by calculating the deviation of (actual-planned) for both inputs
and outputs and then summing the values in the respective planned, actual, and deviation
rows. The initial backlog (at the beginning of period 1) is 30 units. Subsequent backlogs
are calculated by subtracting each period's actual output from the sum of its actual input
and previous backlog.
 Period 1 2 3 4 Total
Planned Input 60 65 75 75 275
Actual Input 60 60 65 65 250
Deviation 0 -5 -10 -10 -25
Planned Output 75 75 75 75 300
Actual Output 70 70 65 65 270
Deviation -5 -5 -10 -10 -30
Backlog 30 20 10 10 10

Completed Input/Output Report

The completed input/output report shows that work center 5 did not process all the jobs
that were available during the four periods; therefore, the desired output rate was not
achieved. This can be attributed to a lower-than-expected input of work from feeding work
centers. The I/O reports from those work centers need to be examined to locate the source
of the problem.

Input/output control provides the information necessary to regulate the flow of work to and
from a network of work centers. Increasing the capacity of a work center that is processing
all the work available to it will not increase output. The source of the problem needs to be
identified. Excessive queues, or backlogs, are one indication that bottlenecks exist. To
alleviate bottleneck work centers, the problem causing the backlog can be worked on, the
capacity of the work center can be adjusted, or input to the work center can be reduced.
Increasing the input to a bottleneck work center will not increase the center's output. It will
merely clog the system further and create longer queues of work-in-process.

5.5 Problems in capacity planning

There is no standard terminology:

Every vendor has a different definition of capacity management, capacity planning, sizing,
tuning, and so on. Some vendors use the term capacity management to include both capacity
planning and tuning. Others use it to denote tuning only.

There is no standard definition of capacity:

One definition is in terms of maximum throughput. Jobs per second, transactions per second
(TPS), Instructions per second (MIPS) or bits per second. Another definition: Maximum
number of users that the system can support while meeting a specified performance
objective.

Capacity planning is expressed in workload units. Users, sessions, tasks, activities, programs,
jobs, accounts, projects and so on.

There are a number different capacities for the same system:

Nominal capacity, usable capacity, and knee capacity. Other terms: practical capacity (usable
capacity) and theoretical capacity (nominal capacity)

There is no standard workload unit:

In case of users or sessions it is difficult to characterize the workload that varies from one
environment to another environment. So workload independent measures such as MIPS are
still popular for forecasting.

Forecasting future applications is difficult:

Most of the forecasting is based on the assumption that the future trend will be similar to the
past. This assumption is violated if new technology suddenly emerges.

There is no uniformity among systems from different vendors:

The same workload takes different amounts of resources on different systems.

This requires developing a vendor independent benchmark and running it on different
systems.

Model inputs cannot always be measured:

Simulation and analytical models are used to predict the performance under different
alternatives.
Sometimes the inputs required for the model are not accurately measurable. Think time is
commonly used in analytical models, but it is impossible to measure think time.

Validating model projections is difficult:

There are two types of model model validations: baseline validation and projection
validation.

Baseline validation requires that using the current workload and configuration in the model
and verifying that the model output matches the observed system performance.

Projection validation, requires changing the workload and configuration and verifying that
the model output matches the changed real system performance are rarely performed so the
model for capacity planning is suspect.

Distributed environments are too complex to model:

Initial computer systems consisted of only a few components. Justifying the cost of each
component was simple. With distributed environments of today the system consists of a large
number of semi-autonomous clients, servers and network links, and I/O devices. Workstation
usage is very different from others and interactions are rather complex. It is difficult to justify
the cost of each component.

Performance is only a small part of the capacity planning problem:

The key issue in capacity planning is cost. Cost of the equipment, software, installation,
maintenance, personnel, floor space, power, and climate control (cooling, humidity control).
Performance modeling helps only in sizing the equipment.

As the cost of computer hardware is declining, the other costs have become a major
consideration in cap city planning.